Harris et al. (2025) Global observations of land-atmosphere interactions during flash drought

Identification

• Journal: Hydrology and earth system sciences
• Year: 2025
• Date: 2025-12-02
• Authors: Bethan L. Harris, Christopher M. Taylor, Wouter Dorigo, Ruxandra-Maria Zotta, Darren Ghent, Iván Noguera
• DOI: 10.5194/hess-29-6917-2025

Research Groups

• National Centre for Earth Observation, UK Centre for Ecology & Hydrology, Wallingford, UK
• Department of Geodesy and Geoinformation, TU Wien, Vienna, Austria
• National Centre for Earth Observation, University of Leicester, Leicester, UK

Short Summary

This study uses global satellite observations to investigate land-atmosphere coupling processes during flash droughts from 2000–2020, revealing that precursor land surface conditions significantly influence drought intensity and associated heat extremes, offering valuable information for subseasonal-to-seasonal (S2S) forecasts.

Objective

• To understand the land-atmosphere coupling processes, specifically soil moisture-evapotranspiration feedbacks, that are critical to flash drought development, in order to evaluate and improve subseasonal-to-seasonal (S2S) forecasts.
• To explore the evolution of the surface energy balance during flash droughts using global long-term daily satellite observations.
• To investigate differences between flash droughts with stronger and weaker land-atmosphere coupling and assess feedbacks from the land surface to near-surface air temperatures.
• To ascertain which globally observable land surface variables provide useful information on flash drought impacts at S2S timescales.

Study Configuration

• Spatial Scale: Global, with analysis at 0.25° pixel resolution and sensitivity analysis at 2.5° grid box resolution. Specific focus on three semi-arid regions in Africa.
• Temporal Scale: 2000–2020 for flash drought identification and analysis. Flash droughts are defined by rapid intensification on subseasonal-to-seasonal (2 weeks–2 months) timescales. Precursor conditions are investigated up to 4 months before drought onset. Daily resolution for most datasets.

Methodology and Data

• Models used:
  • Global Land Evaporation Amsterdam Model (GLEAM) v4.2a: Used to compute latent heat flux and root-zone soil moisture (RZSM), driven by satellite observations and reanalysis.
  • ERA5 reanalysis: Used for 2 m air temperature (T2m), 10 m wind speed, evapotranspiration (ET), and potential evapotranspiration (PET).
• Data sources:
  • ESA Climate Change Initiative (CCI) Soil Moisture combined active/passive microwave product v08.1: Daily surface soil moisture (SSM) at 0.25° resolution (2000–2020).
  • NASA’s Clouds and the Earth’s Radiant Energy System (CERES) daily synoptic product: Net radiation at the surface (Rn) and surface downwelling shortwave radiation (March 2000–2020).
  • ESA CCI Land Surface Temperature project: All-weather MicroWave Land Surface Temperature (MW-LST) v2.33 (0.25° resolution) and MODIS Aqua L3C product v4.00 (0.01° resolution) for Land Surface Temperature (LST) (MW-LST: 2000–2020; MODIS Aqua: July 2002–2020).
  • Integrated Multi-satellitE Retrievals for GPM (IMERG) V06: Daily precipitation data (June 2000–2020).
  • Vegetation Optical Depth Climate Archive (VODCA) v2: Daily combined CXKu-band Vegetation Optical Depth (VOD) (2000–2020).
  • Downscaled GOME-2 Solar Induced Fluorescence (SIF) observations (SIF-JJ, SIF-PK): 8-daily SIF (2007–2018).
  • ESA CCI and Copernicus Climate Change Service (C3S) annual land cover maps: Used for land cover classification.
  • Flash drought identification based on daily standardized anomalies in ESA CCI SM, with a 7-day rolling window for climatological mean/standard deviation.
  • Surface sensible heat flux (H) anomalies are proxied by standardized anomalies of the difference between LST and T2m (ΔT = LST - T2m).
  • Flash droughts are categorized into quartiles based on the maximum ΔT anomaly during the event to assess coupling strength.
  • Sensitivity of variables (VOD, T2m, latent heat flux) to flash drought is calculated as the change in the variable's anomaly divided by the change in SSM anomaly.
  • Statistical testing for T2m peak anomalies uses the Mann-Whitney U test with a cluster bootstrapping method.

Main Results

• During flash drought onset, net radiation (Rn) and surface sensible heat flux (ΔT) increase. Latent heat flux behavior varies by land cover, often decreasing as water limitation is reached, particularly in shrubland.
• After onset, ΔT continues to increase even as Rn declines, indicating a shift towards a water-limited soil regime where the surface energy balance partitions more towards sensible heat flux.
• Flash droughts with stronger evaporative stress (higher peak ΔT) are associated with larger decreases in latent heat flux, drier root-zone soil moisture, and stronger negative evaporative stress ratios.
• Events with stronger evaporative stress exhibit perturbed surface energy budgets (elevated ΔT, suppressed latent heat flux) and below-average root-zone soil moisture for up to 4 months both before and after drought onset, highlighting the importance of precursor land conditions.
• Stronger flash droughts lead to more severe impacts on vegetation, as indicated by larger decreases in Vegetation Optical Depth (VOD) and Solar Induced Fluorescence (SIF). VOD shows more prolonged effects than SIF.
• Semi-arid regions show the highest sensitivity of VOD and near-surface air temperature (T2m) to flash droughts, consistent with water-limited conditions promoting strong land-atmosphere coupling.
• In West Africa (June-August) and East Africa (March-May), lower VOD 1–2 months before flash drought onset is significantly linked to a higher probability of extreme peak T2m anomalies during the event. For example, in West African summer, 27% of flash droughts with precursor VOD anomalies in the lowest quartile experience a peak T2m anomaly > 1.5 standard deviations, compared to 12% for those in the highest quartile.
• Higher peak sensible heat flux during flash droughts feeds back to increase near-surface air temperatures in the studied African regions.

Contributions

• Provides the first global observational analysis of land-atmosphere interactions during flash droughts, minimizing reliance on reanalysis data with known biases in land-atmosphere coupling.
• Demonstrates a physically consistent picture of the surface energy budget evolution during flash droughts using a diverse set of independent global satellite-based observational products (e.g., SSM, LST, VOD, SIF, Rn).
• Quantifies the significant influence of precursor land surface conditions (up to 4 months prior) on flash drought intensity and associated heat extremes, establishing a basis for subseasonal-to-seasonal (S2S) predictability.
• Highlights the practical utility of globally observable land surface conditions, such as Vegetation Optical Depth (VOD), as a proxy for root-zone soil moisture and a predictor for heat extremes in data-sparse regions like Africa, with a lead time of 1–2 months.
• Motivates further development and evaluation of land-atmosphere coupling, including dynamic vegetation models and critical soil moisture thresholds, in S2S forecasting models to improve flash drought prediction skill.

Funding

• European Space Agency Climate Change Initiative Research Fellowship (contract number 4000141837/23/I/DT-lr)
• Natural Environment Research Council’s support of the National Centre for Earth Observation, via the project CPEO (Constraining Coupled Carbon & Water Cycle Processes with Earth Observation), Grant No. NE/X006328/1
• ESA CCI Soil Moisture project (ESRIN Contract No: 4000126684/19/I-NB)

Citation

@article{Harris2025Global,
  author = {Harris, Bethan L. and Taylor, Christopher M. and Dorigo, Wouter and Zotta, Ruxandra-Maria and Ghent, Darren and Noguera, Iván},
  title = {Global observations of land-atmosphere interactions during flash drought},
  journal = {Hydrology and earth system sciences},
  year = {2025},
  doi = {10.5194/hess-29-6917-2025},
  url = {https://doi.org/10.5194/hess-29-6917-2025}
}